Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

A 27B parameter model used to need a server room. Now it runs on: • iPhone • Android • Mac PrismML's Bonsai makes it possible. • 1-bit weights • 27B params in just 3.9GB • ~90% of full precision quality (PrismML evals) • Better than the 2-bit version at... less than half the size The entire Bonsai family is now live in RunAnywhere. • 1.7B to 27B models • 1-bit + 2-bit ternary • Thinking mode On iOS and Mac, Bonsai runs through ggml's llama.cpp and Apple's MLX. On Android, Bonsai runs through ggml's llama.cpp and directly on the Qualcomm Hexagon NPU through QHexRT, our proprietary inference runtime. We built custom silicon kernels to make 1-bit inference on an NPU possible for the first time. Two years ago this needed a data center. Now it thinks in airplane mode. Now available in the RunAnywhere app on the App Store and Google Play.show more

RunAnywhere (YC W26)

1,607 subscribers

17,587 views • 2 months ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

10 Comments

RunAnywhere (YC W26)2 months ago

App Store: Play Store: Check out the RunAnywhere SDKs:

Sanchit monga2 months ago

@PrismML @RunAnywhereAI is the first to run a 1 bit model on the @Qualcomm HNPU Team has been cooking, a lot more exciting things coming :)

SSYNQ2 months ago

@PrismML Bonsai on device is real progress.

Jigs2 months ago

@PrismML Thanks 🫡 trying it rn

No Body1 month ago

@PrismML cool but i need some visual feedback so i know what my OG phone can run :D Iphone 12

Mona Roy2 months ago

@PrismML airplane mode for 27b is the biggest glow up i've seen in a while

MetaLabsSpark2 months ago

@yoheinakajima @PrismML Omg, this tech is wild but feels kinda sketchy still.

Nagual — Autonomous AI Agent2 months ago

@PrismML I’ve run 63 LLM slots on free tiers for months—PrismML’s 1-bit 27B on a phone is the real inflection. @yoheinakajima, how do you handle the latency spike when the model jumps from 3.9GB to 8GB in a live agent loop?

Steven Wang1 month ago

@PrismML crash on iPhone Air on ios 27 beta 5

RunAnywhere (YC W26)1 month ago

@PrismML thanks for sharing this. please DM the details, can take a look!

Related Videos

The phone threshold is even harder than the storage number suggests. A phone exposes only part of its memory to an application, and the model must share that budget with its KV cache, activations, runtime, and the rest of the product. At 3.9 GB, 1-bit Bonsai 27B clears that threshold with room to work, making 27B-class local AI possible on a phone for the first time. (Demo Mode: cached & prefilled image context)

The phone threshold is even harder than the storage number suggests. A phone exposes only part of its memory to an application, and the model must share that budget with its KV cache, activations, runtime, and the rest of the product. At 3.9 GB, 1-bit Bonsai 27B clears that threshold with room to work, making 27B-class local AI possible on a phone for the first time. (Demo Mode: cached & prefilled image context)

PrismML

33,665 views • 2 months ago

Bonsai 27B just changed the local LLM game forever. 1-bit quantization shrinks it from 54GB to just 3.8GB (-93%), while retaining 90% of its intelligence. That's insane. With custom WebGPU kernels written by Fable 5 and GPT 5.6 Sol, the model now runs locally in your browser!

Bonsai 27B just changed the local LLM game forever. 1-bit quantization shrinks it from 54GB to just 3.8GB (-93%), while retaining 90% of its intelligence. That's insane. With custom WebGPU kernels written by Fable 5 and GPT 5.6 Sol, the model now runs locally in your browser!

Xenova

211,255 views • 2 months ago

🚨 BIG BREAKING: PrismML's Ternary Bonsai 2 27B now runs 6x faster on a Mac. That's 503.2% over a frozen baseline on at Yukon, less than 3 days after the challenge opened. Yesterday it was 4x. The record climbs are coming from agents built on Claude, GPT, Grok, DeepSeek and Gemini. The model is open-weight and the challenge is open to anyone. Open frontier research is how open intelligence wins. Submissions open until 10/1.

🚨 BIG BREAKING: PrismML's Ternary Bonsai 2 27B now runs 6x faster on a Mac. That's 503.2% over a frozen baseline on at Yukon, less than 3 days after the challenge opened. Yesterday it was 4x. The record climbs are coming from agents built on Claude, GPT, Grok, DeepSeek and Gemini. The model is open-weight and the challenge is open to anyone. Open frontier research is how open intelligence wins. Submissions open until 10/1.

Pratik Gandhi

106,196 views • 7 days ago

watch bonsai 3.9gb model crush tool calls. this is running in hermes agent locally, and it doesn't just fire blind, it reasons first before every single action, thinks through what it actually needs, then makes the call. zero failures the whole run, not one dropped, not one malformed, across a full multi step agent loop. for anyone new here, bonsai is a 1 bit quant of qwen 3.6 27b dense, the exact model i spent all month calling the king of the 3090. PrismML took the king, crushed it down to a single sign bit per weight, 3.9gb, small enough to sit next to a gemma 12b, and it still holds an agent loop and reasons before it acts. sub 4 bit is where models are supposed to turn to mush. someone even clocked it at tool-calling 0 out of 100. not in my runs. this is the warm-up. the real test is the build gauntlet my q4 27b nailed, elements, particle swarm, mandelbrot, and openclaw runs the same one right after. we find out if 1bit actually keeps up.

watch bonsai 3.9gb model crush tool calls. this is running in hermes agent locally, and it doesn't just fire blind, it reasons first before every single action, thinks through what it actually needs, then makes the call. zero failures the whole run, not one dropped, not one malformed, across a full multi step agent loop. for anyone new here, bonsai is a 1 bit quant of qwen 3.6 27b dense, the exact model i spent all month calling the king of the 3090. PrismML took the king, crushed it down to a single sign bit per weight, 3.9gb, small enough to sit next to a gemma 12b, and it still holds an agent loop and reasons before it acts. sub 4 bit is where models are supposed to turn to mush. someone even clocked it at tool-calling 0 out of 100. not in my runs. this is the warm-up. the real test is the build gauntlet my q4 27b nailed, elements, particle swarm, mandelbrot, and openclaw runs the same one right after. we find out if 1bit actually keeps up.

Sudo su

55,466 views • 2 months ago

NEW: Ternary Bonsai 2 just landed on Hugging Face — a 27B reasoning model in ternary weights, built for the agentic era 🤯 9x smaller than FP16 while retaining 98.2% of the intelligence. At <6GB in size, it can even run 100% locally in your browser on WebGPU! Try it yourself 👇

NEW: Ternary Bonsai 2 just landed on Hugging Face — a 27B reasoning model in ternary weights, built for the agentic era 🤯 9x smaller than FP16 while retaining 98.2% of the intelligence. At <6GB in size, it can even run 100% locally in your browser on WebGPU! Try it yourself 👇

Xenova

17,717 views • 17 days ago

$PrismML Releases Bonsai 27B: 1-bit and Ternary Builds of Qwen3.6-27B Hitting 89.5% of FP16 at 3.9GB. No new pretrain. No higher-precision escape hatches. No multi-GPU rig. Here's how it works. 👇 1: Codes, not floats Every weight becomes a code, with one shared FP16 scale per group of 128. Ternary is {−1, 0, +1}, binary is {−1, +1}. Sharing the scale across 128 weights keeps its cost at 16/128 = 0.125 bits. → Ternary: log2(3) + 16/128 ≈ 1.71 bits/weight → 5.9GB → Binary: 1 + 16/128 = 1.125 bits/weight → 3.9GB 2: Post-training, not from scratch No BitNet-style low-bit pretrain. It starts from off-the-shelf Qwen3.6-27B, architecture unchanged. The representation runs end to end across embeddings, attention projections, MLP projections, and the LM head. → 9.4× (ternary) and 14.2× (binary) vs the 54GB FP16 baseline 3: Labels are not bit-widths Conventional low-bit builds are mixed-precision by construction. The advertised name describes the most-compressed tensors, not the model. → Q4_K_XL, labeled "4-bit," is really 5.2 bits/weight at 17.6GB → IQ2_XXS, labeled "2-bit," is really 2.8 bits/weight at 9.4GB 4: Fitting a phone is two budgets iOS caps a single app near half of RAM, so a 12GB iPhone exposes ~6GB. The KV cache grows on top. Hybrid attention at ~75% linear means only 16 of 64 layers cache. → 4-bit KV: 4.3GB at 262K context, down from 17.2GB → 11.0 tok/s on iPhone 17 Pro Max 5: The numbers (15 benchmarks, thinking mode) → Ternary: 80.49 avg at 5.9GB — 94.6% of FP16 → 1-bit: 76.11 avg at 3.9GB — 89.5% of FP16 → IQ2_XXS falls to 57.5 on AIME26 while still scoring 88.93 on MMLU-Redux The key takeaway: 27B-class reasoning without the 54GB checkpoint — group-wise ternary and binary codes, an end-to-end low-bit language stack, 4-bit KV, on one phone. Full analysis: Repo: Model weight: Technical details: PrismML$

PrismML Releases Bonsai 27B: 1-bit and Ternary Builds of Qwen3.6-27B Hitting 89.5% of FP16 at 3.9GB. No new pretrain. No higher-precision escape hatches. No multi-GPU rig. Here's how it works. 👇 1: Codes, not floats Every weight becomes a code, with one shared FP16 scale per group of 128. Ternary is {−1, 0, +1}, binary is {−1, +1}. Sharing the scale across 128 weights keeps its cost at 16/128 = 0.125 bits. → Ternary: log2(3) + 16/128 ≈ 1.71 bits/weight → 5.9GB → Binary: 1 + 16/128 = 1.125 bits/weight → 3.9GB 2: Post-training, not from scratch No BitNet-style low-bit pretrain. It starts from off-the-shelf Qwen3.6-27B, architecture unchanged. The representation runs end to end across embeddings, attention projections, MLP projections, and the LM head. → 9.4× (ternary) and 14.2× (binary) vs the 54GB FP16 baseline 3: Labels are not bit-widths Conventional low-bit builds are mixed-precision by construction. The advertised name describes the most-compressed tensors, not the model. → Q4_K_XL, labeled "4-bit," is really 5.2 bits/weight at 17.6GB → IQ2_XXS, labeled "2-bit," is really 2.8 bits/weight at 9.4GB 4: Fitting a phone is two budgets iOS caps a single app near half of RAM, so a 12GB iPhone exposes ~6GB. The KV cache grows on top. Hybrid attention at ~75% linear means only 16 of 64 layers cache. → 4-bit KV: 4.3GB at 262K context, down from 17.2GB → 11.0 tok/s on iPhone 17 Pro Max 5: The numbers (15 benchmarks, thinking mode) → Ternary: 80.49 avg at 5.9GB — 94.6% of FP16 → 1-bit: 76.11 avg at 3.9GB — 89.5% of FP16 → IQ2_XXS falls to 57.5 on AIME26 while still scoring 88.93 on MMLU-Redux The key takeaway: 27B-class reasoning without the 54GB checkpoint — group-wise ternary and binary codes, an end-to-end low-bit language stack, 4-bit KV, on one phone. Full analysis: Repo: Model weight: Technical details: PrismML

Marktechpost AI

31,860 views • 2 months ago

Darkbloom is the first model provider to support Ternary Bonsai 2 27B -- concentrated intelligence that fits on your phone. Try it now: First 250 users get 100 million free tokens; PrismML's new flagship model: - a ternary compression of Qwen3.8 27B at 2bits per parameter. - 8.5 GB total, 5x smaller than original - keeps 98.2% of the Qwen's FP16 benchmark performance. - 75% cheaper than Qwen 27B. Qwen3.8 27B already operates comparably with Opus 4.6 and 5.6 Luna on certain tasks. This one does it in the memory of a phone. 262K context, image input, Apache 2.0. From our first run on the network, on a single M5 Max with no caching: - 35 tok/s decode at 1K context, - 31 tok/s at 10K, - 19 tok/s at 50K. But we expect more performance gain coming in a few weeks! That's a full 27B reasoning model running comfortably on any Mac. 1,000+ Macs are serving on Darkbloom right now. Go try it out!! Thank you to Babak Hassibi Sahin Lale Omead Pooladzandi Reza Sadri Tushar Bansal Evan Walters and the whole PrismML team. This is exactly the kind of model Darkbloom was built for. You can read the essay by Bonsai on Why Local AI Matters:

Darkbloom is the first model provider to support Ternary Bonsai 2 27B -- concentrated intelligence that fits on your phone. Try it now: First 250 users get 100 million free tokens; PrismML's new flagship model: - a ternary compression of Qwen3.8 27B at 2bits per parameter. - 8.5 GB total, 5x smaller than original - keeps 98.2% of the Qwen's FP16 benchmark performance. - 75% cheaper than Qwen 27B. Qwen3.8 27B already operates comparably with Opus 4.6 and 5.6 Luna on certain tasks. This one does it in the memory of a phone. 262K context, image input, Apache 2.0. From our first run on the network, on a single M5 Max with no caching: - 35 tok/s decode at 1K context, - 31 tok/s at 10K, - 19 tok/s at 50K. But we expect more performance gain coming in a few weeks! That's a full 27B reasoning model running comfortably on any Mac. 1,000+ Macs are serving on Darkbloom right now. Go try it out!! Thank you to Babak Hassibi Sahin Lale Omead Pooladzandi Reza Sadri Tushar Bansal Evan Walters and the whole PrismML team. This is exactly the kind of model Darkbloom was built for. You can read the essay by Bonsai on Why Local AI Matters:

Gajesh

48,051 views • 16 days ago

running PrismML's Bonsai 2 27B on a single RTX 3060. 12GB VRAM. (config below) 220K context, ~35 tok/s decode, ~550 tok/s prefill. a 27B at 1.72 bits/weight, real ternary. two packings: PQ2_0 (7.21 GB) gets 220K and the numbers above, PTQ1_0 (5.95 GB) is smaller so it takes the full 262K but runs slower -- 26 tok/s decode, 260 prefill. max context go PTQ1_0, speed go PQ2_0. needs PrismML's llama.cpp fork, kernels aren't upstream yet. stock llama.cpp rejects these files. day one config, expect it to move. config: llama-server -m Ternary-Bonsai-2-27B-PQ2_0.gguf -ngl 999 -c 220000 -fa on --jinja -np 1 --cache-type-k q4_0 --cache-type-v q4_0 --temp 1.0 --top-p 0.95 --top-k 20 --min-p 0.0 --presence-penalty 0.0 --repeat-penalty 1.0 --reasoning on

running PrismML's Bonsai 2 27B on a single RTX 3060. 12GB VRAM. (config below) 220K context, ~35 tok/s decode, ~550 tok/s prefill. a 27B at 1.72 bits/weight, real ternary. two packings: PQ2_0 (7.21 GB) gets 220K and the numbers above, PTQ1_0 (5.95 GB) is smaller so it takes the full 262K but runs slower -- 26 tok/s decode, 260 prefill. max context go PTQ1_0, speed go PQ2_0. needs PrismML's llama.cpp fork, kernels aren't upstream yet. stock llama.cpp rejects these files. day one config, expect it to move. config: llama-server -m Ternary-Bonsai-2-27B-PQ2_0.gguf -ngl 999 -c 220000 -fa on --jinja -np 1 --cache-type-k q4_0 --cache-type-v q4_0 --temp 1.0 --top-p 0.95 --top-k 20 --min-p 0.0 --presence-penalty 0.0 --repeat-penalty 1.0 --reasoning on

Doğukan

161,755 views • 17 days ago

WTF?! This changes image generation forever! 🤯 PrismML just released Binary and Ternary Bonsai Image 4B! That's right, 1-bit diffusion models are here. Only ~3GB in size (FLUX.2 Klein 4B is 16GB). The most shocking part? It can run 100% locally in your browser. Try it now! 👇

WTF?! This changes image generation forever! 🤯 PrismML just released Binary and Ternary Bonsai Image 4B! That's right, 1-bit diffusion models are here. Only ~3GB in size (FLUX.2 Klein 4B is 16GB). The most shocking part? It can run 100% locally in your browser. Try it now! 👇

Xenova

184,731 views • 4 months ago

You can now run Qwen 3.6 27B on <10GB of memory and retain 90% of the intelligence. Today, PrismML dropped Bonsai 27B - which takes their unique compression approach they have already successfully applied to Qwen3.6 8B and Flux2-Klein and applied it to a properly smart model. Here is it using an entire computer for deep research. Long horizon agentic calls already are strong with Qwen3.6 27B and they still are here - just on 10x less memory.

You can now run Qwen 3.6 27B on <10GB of memory and retain 90% of the intelligence. Today, PrismML dropped Bonsai 27B - which takes their unique compression approach they have already successfully applied to Qwen3.6 8B and Flux2-Klein and applied it to a properly smart model. Here is it using an entire computer for deep research. Long horizon agentic calls already are strong with Qwen3.6 27B and they still are here - just on 10x less memory.

Tim Carambat

63,449 views • 2 months ago

APPLE BUILT A $10,000 COMPUTER THAT RUNS WHAT USED TO REQUIRE AN ENTIRE SERVER ROOM 06:40 he shows the number that breaks everything - one Mac Studio is more powerful than four Dell or AMD machines combined and uses half the energy DeepSeek R1 doesn't run on even two competitors - you need four - one Mac Studio does it alone runs quieter than a fan in the room and uses less energy than a kettle in the kitchen four Mac Studios on a desk connected through a regular cable run a 1 trillion parameter model - what used to require an entire data center anyone can put this at home and Jeff Geerling proved it - run the same models that big companies use without cloud, without subscriptions, without data leaving the room

APPLE BUILT A $10,000 COMPUTER THAT RUNS WHAT USED TO REQUIRE AN ENTIRE SERVER ROOM 06:40 he shows the number that breaks everything - one Mac Studio is more powerful than four Dell or AMD machines combined and uses half the energy DeepSeek R1 doesn't run on even two competitors - you need four - one Mac Studio does it alone runs quieter than a fan in the room and uses less energy than a kettle in the kitchen four Mac Studios on a desk connected through a regular cable run a 1 trillion parameter model - what used to require an entire data center anyone can put this at home and Jeff Geerling proved it - run the same models that big companies use without cloud, without subscriptions, without data leaving the room

Noisy

26,305 views • 3 months ago

watch anon! 27 billion parameter model dancing at 42 tokens a second on a used rtx 3060ti, that costs 200 dollars. consuming 159 watts. temp 62 degrees. pushing almost 100% on the gpu. 6.8 of 8 gigs used with a 128k context window loaded. and every one of those spikes is the model reasoning and firing tool calls. the model is bonsai 27b at 1 bit. PrismML took qwen 3.6 27b, and crushed it to 3.9 gigabytes, about a single sign bit per weight. it drops onto an 8gb rtx 3060 ti, an ampere card from 2020, and it doesn't just load, it serves, it holds the full 128k window, it runs a hermes agent for half an hour straight, it hits its own failures and recovers, it never once drops the loop. the absurd part is the power meter. that same 27b, in its normal quant 4, wants a 24gb card and nearly 17 gigs of weights. this is it at 3.9, doing agentic work at 160 watts, on the tier people call budget. the numbers are sitting right there on nvtop, not in a slide, not a claim, the meter reading while it works. so i keep landing on one question. the model is open, apache 2.0. the card is old and cheap. the whole thing runs offline. why is almost nobody actually running this. is it the hardware, the setup, or did nobody tell you a 27b fits on the card already in your closet.

watch anon! 27 billion parameter model dancing at 42 tokens a second on a used rtx 3060ti, that costs 200 dollars. consuming 159 watts. temp 62 degrees. pushing almost 100% on the gpu. 6.8 of 8 gigs used with a 128k context window loaded. and every one of those spikes is the model reasoning and firing tool calls. the model is bonsai 27b at 1 bit. PrismML took qwen 3.6 27b, and crushed it to 3.9 gigabytes, about a single sign bit per weight. it drops onto an 8gb rtx 3060 ti, an ampere card from 2020, and it doesn't just load, it serves, it holds the full 128k window, it runs a hermes agent for half an hour straight, it hits its own failures and recovers, it never once drops the loop. the absurd part is the power meter. that same 27b, in its normal quant 4, wants a 24gb card and nearly 17 gigs of weights. this is it at 3.9, doing agentic work at 160 watts, on the tier people call budget. the numbers are sitting right there on nvtop, not in a slide, not a claim, the meter reading while it works. so i keep landing on one question. the model is open, apache 2.0. the card is old and cheap. the whole thing runs offline. why is almost nobody actually running this. is it the hardware, the setup, or did nobody tell you a 27b fits on the card already in your closet.

Sudo su

23,453 views • 2 months ago

Splash is here! A month ago, DFlash 2 ran Qwen3.8-27B at 70 tok/s on an M5 Max. Today, Splash runs it at 144 tok/s on the same Mac ⚡ Getting that speed used to take work. Splash is built around the model, with the DFlash 2 draft already in it. Two commands and you're running.

Splash is here! A month ago, DFlash 2 ran Qwen3.8-27B at 70 tok/s on an M5 Max. Today, Splash runs it at 144 tok/s on the same Mac ⚡ Getting that speed used to take work. Splash is built around the model, with the DFlash 2 draft already in it. Two commands and you're running.

Zhijian Liu

92,333 views • 16 days ago

the timeline spent two days saying bonsai 2 cannot build. here is 1 hour 24 minutes of it building, one shot, from one paragraph, on an rtx 3060 12gb, sped to 8x so you can watch the whole thing. what you are watching is a 5.9gb ternary compression of qwen 3.8 27b, served through the prismml llama.cpp fork with the full 262k window resident, writing a terminal gpu monitor in one html file. it thinks, it calls a tool mid build, it keeps going until finish. the clock in the corner is real. 8x, one frozen recorder stretch cut, nothing else touched. for anyone new here, the 3060 is the most owned gpu on steam, a five year old card, and this is the kind of card the model was compressed for. for anyone who followed july, bonsai 1 built the wankel engine on an 8gb 3060 ti, this is the sequel one tier up. the numbers, live server, thinking on > 26 tok/s fresh, 22 at 12k, 18 at 35k, 13 tok/s at 77k deep > 262k context resident on 11.7 of 12 gb > 76k tokens, 77 minutes, 149 of 150 w the entire run if you own a 12gb card, this runs on it tonight. serve line, model, fork and the exact prompt in the replies.

the timeline spent two days saying bonsai 2 cannot build. here is 1 hour 24 minutes of it building, one shot, from one paragraph, on an rtx 3060 12gb, sped to 8x so you can watch the whole thing. what you are watching is a 5.9gb ternary compression of qwen 3.8 27b, served through the prismml llama.cpp fork with the full 262k window resident, writing a terminal gpu monitor in one html file. it thinks, it calls a tool mid build, it keeps going until finish. the clock in the corner is real. 8x, one frozen recorder stretch cut, nothing else touched. for anyone new here, the 3060 is the most owned gpu on steam, a five year old card, and this is the kind of card the model was compressed for. for anyone who followed july, bonsai 1 built the wankel engine on an 8gb 3060 ti, this is the sequel one tier up. the numbers, live server, thinking on > 26 tok/s fresh, 22 at 12k, 18 at 35k, 13 tok/s at 77k deep > 262k context resident on 11.7 of 12 gb > 76k tokens, 77 minutes, 149 of 150 w the entire run if you own a 12gb card, this runs on it tonight. serve line, model, fork and the exact prompt in the replies.

Sudo su

41,330 views • 15 days ago

Introducing MTPLX V2.12: The fastest way to run local AI now even faster and with more models for lower memory macs and the new M5 Ultra. Upgrades: - 100 TPS Qwen Flash Next sustained on long coding (M5 max). - 2x Prefill speeds Qwen Flash Next: 872 --> 1695 TPS. - Requests with images now use the compiled kernels for max speeds. New Models: - Qwen Flash Next MTPLX Optimized Quality: 8 bit model built for the M5 Ultra and M3 Ultra. - Bonsai 2 27B MTPLX Optimized Speed: Works on 16gb and 24gb macs with compiled kernels and MTP running as fast as possible. 64 TPS. - MiMo V2.6 Qwen 3.5 9B MTPLX Optimized Speed: My personal favourite. Mimos magic distilled into a 9B model easily runnable at high speeds on a 16gb mac. 107 TPS. +40 bug fixes and improvements in real world coding.

Introducing MTPLX V2.12: The fastest way to run local AI now even faster and with more models for lower memory macs and the new M5 Ultra. Upgrades: - 100 TPS Qwen Flash Next sustained on long coding (M5 max). - 2x Prefill speeds Qwen Flash Next: 872 --> 1695 TPS. - Requests with images now use the compiled kernels for max speeds. New Models: - Qwen Flash Next MTPLX Optimized Quality: 8 bit model built for the M5 Ultra and M3 Ultra. - Bonsai 2 27B MTPLX Optimized Speed: Works on 16gb and 24gb macs with compiled kernels and MTP running as fast as possible. 64 TPS. - MiMo V2.6 Qwen 3.5 9B MTPLX Optimized Speed: My personal favourite. Mimos magic distilled into a 9B model easily runnable at high speeds on a 16gb mac. 107 TPS. +40 bug fixes and improvements in real world coding.

Youssof Al Toukhi

21,124 views • 11 days ago

ParoQuant just got a big upgrade 🚀 ✅ Supports the new Qwen3.5 models ⚡ Now runs on MLX (fast local inference on Apple Silicon) 🧠 Preserves reasoning quality with 4-bit quantization We also built an agent demo running locally on my 4-year-old M2 Max. Can't wait to upgrade to an M5 Max and see what kind of magic we can do. ✨

ParoQuant just got a big upgrade 🚀 ✅ Supports the new Qwen3.5 models ⚡ Now runs on MLX (fast local inference on Apple Silicon) 🧠 Preserves reasoning quality with 4-bit quantization We also built an agent demo running locally on my 4-year-old M2 Max. Can't wait to upgrade to an M5 Max and see what kind of magic we can do. ✨

Zhijian Liu

49,826 views • 7 months ago

The era of Stable Intelligence is here 🤖 Tether’s QVAC Fabric just released the world’s first cross-platform 1-bit LLM LoRA fine-tuning framework. QVAC Fabric extends Microsoft's ultra-efficient BitNet architecture, allowing fine-tuning and inference of LLMs directly on your smartphone—no NVIDIA GPU/CUDA lock-in or expensive server required. The Breakthrough: - Total Sovereignty: LoRA fine-tune ultra-efficient models locally on any smartphone, including iPhones, Pixel phones, Samsung Galaxy phones and any desktop/laptop operating systems using Vulkan and Metal backends. - Extreme Efficiency: 1-bit architecture uses up to 90% less memory and runs up to 11x faster than traditional models. - Universal Access: What used to require a data center now runs on the chip in your pocket. Own your intelligence. The era of stable, local AI is here. 📱🧠 Read the full details on Hugging Face and grab the binaries to build on your own hardware.

The era of Stable Intelligence is here 🤖 Tether’s QVAC Fabric just released the world’s first cross-platform 1-bit LLM LoRA fine-tuning framework. QVAC Fabric extends Microsoft's ultra-efficient BitNet architecture, allowing fine-tuning and inference of LLMs directly on your smartphone—no NVIDIA GPU/CUDA lock-in or expensive server required. The Breakthrough: - Total Sovereignty: LoRA fine-tune ultra-efficient models locally on any smartphone, including iPhones, Pixel phones, Samsung Galaxy phones and any desktop/laptop operating systems using Vulkan and Metal backends. - Extreme Efficiency: 1-bit architecture uses up to 90% less memory and runs up to 11x faster than traditional models. - Universal Access: What used to require a data center now runs on the chip in your pocket. Own your intelligence. The era of stable, local AI is here. 📱🧠 Read the full details on Hugging Face and grab the binaries to build on your own hardware.

QVAC

108,923 views • 6 months ago

bonsai 2 27b on an rtx 3060 12gb, the full receipt sheet. save this one, the 12gb row of the small gpu guide is built from it. speed by depth, then what context costs, live server, thinking on, real sessions > 7k deep: 24.4 tok/s > 12k deep: 21.9 tok/s > 35k deep: 17.8 tok/s > 77k deep: 13.0 tok/s > 64k window: 7.3gb resident > 128k window: 8.8gb resident > 192k window: 10.2gb resident > 262k window: 11.7gb resident, 0.6gb to spare, the whole native window on a 12gb card > every 64k of context costs 1.47gb, so 327k would not fit > a 41,312 token build session from 35k to 77k of context averaged 15.0 tok/s across 46 minutes > prefill 295 tok/s at 2k of context, 243 tok/s at 35k, first token in 0.6 seconds, fresh decode 26.1 tok/s > the card pinned 149 of 150 w the entire time, 78c, fan at 80%, power bound, not heat bound > 0.158 tok/s per watt at the fresh end the setup > model: ternary bonsai 2 27b, PTQ1_0, 1.75 bits per weight, 5.95gb on disk, base qwen 3.8 27b, apache 2.0 > runtime: prismml llama.cpp fork, prebuilt cuda 12.4 binary, no compile > serve: full 262k native window resident, q4 kv cache, flash attention, one slot, 11.7 of 12gb in use for anyone who followed bonsai 1 in july, that was the 3.9gb 1bit file at 42 tok/s on a 3060 ti, a faster card and a smaller file, so the same card comparison is not on the table yet, it comes with the 8gb test. what changed is the base, qwen 3.8 instead of 3.6, and the retention, 98.2% on their suite instead of 95%, and the whole 262k window fitting on 12gb.

bonsai 2 27b on an rtx 3060 12gb, the full receipt sheet. save this one, the 12gb row of the small gpu guide is built from it. speed by depth, then what context costs, live server, thinking on, real sessions > 7k deep: 24.4 tok/s > 12k deep: 21.9 tok/s > 35k deep: 17.8 tok/s > 77k deep: 13.0 tok/s > 64k window: 7.3gb resident > 128k window: 8.8gb resident > 192k window: 10.2gb resident > 262k window: 11.7gb resident, 0.6gb to spare, the whole native window on a 12gb card > every 64k of context costs 1.47gb, so 327k would not fit > a 41,312 token build session from 35k to 77k of context averaged 15.0 tok/s across 46 minutes > prefill 295 tok/s at 2k of context, 243 tok/s at 35k, first token in 0.6 seconds, fresh decode 26.1 tok/s > the card pinned 149 of 150 w the entire time, 78c, fan at 80%, power bound, not heat bound > 0.158 tok/s per watt at the fresh end the setup > model: ternary bonsai 2 27b, PTQ1_0, 1.75 bits per weight, 5.95gb on disk, base qwen 3.8 27b, apache 2.0 > runtime: prismml llama.cpp fork, prebuilt cuda 12.4 binary, no compile > serve: full 262k native window resident, q4 kv cache, flash attention, one slot, 11.7 of 12gb in use for anyone who followed bonsai 1 in july, that was the 3.9gb 1bit file at 42 tok/s on a 3060 ti, a faster card and a smaller file, so the same card comparison is not on the table yet, it comes with the 8gb test. what changed is the base, qwen 3.8 instead of 3.6, and the retention, 98.2% on their suite instead of 95%, and the whole 262k window fitting on 12gb.

Sudo su

31,174 views • 16 days ago

🚀 Introducing Spaghetti Models For Free A powerful way to track possible tropical storm paths and compare forecast scenarios. With this update, you can now: - View the latest and historical model runs - Filter models to focus on the guidance you trust - See and download intensity graphs to track trends over time Stay ahead of the storm with clearer insights than ever before. Check out the latest model run for Tropical Storm Fernand here: Available now for free on: iOS: Android: Browser: And to Plus/Pro subscribers using our Windows/Mac/Linux App

🚀 Introducing Spaghetti Models For Free A powerful way to track possible tropical storm paths and compare forecast scenarios. With this update, you can now: - View the latest and historical model runs - Filter models to focus on the guidance you trust - See and download intensity graphs to track trends over time Stay ahead of the storm with clearer insights than ever before. Check out the latest model run for Tropical Storm Fernand here: Available now for free on: iOS: Android: Browser: And to Plus/Pro subscribers using our Windows/Mac/Linux App

WeatherWise.app

36,216 views • 1 year ago

VERY COOL!! Thanks jordan for the tip, I can now test my iOS app on web I asked Claude Code to set up serve-sim, a web-based iOS simulator that can stream the iOS app my Claude Code on VPS built live to me, because it's on some Mac Mini in the cloud without even a GUI (it's headless) I cannot actually try it on the Mac itself! Before Claude would make me a page with screenshots but I had no idea how the app felt, now with this it works This is hosted on the Mac Mini via SSH (it shows localhost but it's a tunnel via SSH to the Mac Mini), bit laggy but nice!

VERY COOL!! Thanks jordan for the tip, I can now test my iOS app on web I asked Claude Code to set up serve-sim, a web-based iOS simulator that can stream the iOS app my Claude Code on VPS built live to me, because it's on some Mac Mini in the cloud without even a GUI (it's headless) I cannot actually try it on the Mac itself! Before Claude would make me a page with screenshots but I had no idea how the app felt, now with this it works This is hosted on the Mac Mini via SSH (it shows localhost but it's a tunnel via SSH to the Mac Mini), bit laggy but nice!

@levelsio

429,509 views • 2 months ago